Proceedings of the International MultiConference of Engineers and Computer Scientists 2012 Vol II,
IMECS 2012, March 14 - 16, 2012, Hong Kong
Kinetics of High Temperature Oxidation of
High Carbon Steels in Multi-component Gases
Approximating Industrial Steel Reheat Furnace
Atmospheres
H. T. Abuluwefa
Abstract— The isothermal oxidation behaviour of some
carbon steels under various experimental conditions was
investigated. The composition of oxidation atmosphere used
in this work was chosen to approximate atmospheres found in
industrial steel billet reheat furnaces. In order to study the
effect of adding each gaseous component in these atmospheres
experiments were also conducted in atmospheres leading to
these compositions. Since the most significant oxidation of
the steel occurs at high temperatures, three temperatures of
1000oC, 1100oC and 1200oC were used in the experiments. In
general, it was found that rates of oxidation in free oxygen
atmospheres were higher than the rates of oxidation in
nitrogen-based CO2 and H2O atmospheres. For oxidation in
atmospheres containing higher free oxygen content, the overall
oxidation is unaffected by changing the concentration of H2O
or CO2. However, oxidation rates were found to be
independent of oxygen concentrations above 6 vol.% O2.
Metallographic examination of the oxides layers revealed that
in the presence of free oxygen, the three iron oxides, wustite
FeO, magnetite Fe3O4 and hematite Fe2O3 were all formed.
However, for oxidation in nitrogen-based CO2 and H2O
atmospheres only the wustite phase was present.
Index Terms— Oxidation, Reaction Kinetics, Reheat
Furnaces, Process Modelling.
I. INTRODUCTION
T
HE phenomena of scale formation (oxidation) on steel
surfaces during reheating of steel stock in industrial
reheat furnaces are unavoidable and difficult to control.
Disadvantages of scale formation include metal consumption
and scale fall out in route to rolling units, which requires clean
up and may present environmental problems. Factors which
affect this process include temperature, residence time in the
furnace, furnace atmosphere and steel composition. During
reheating in industrial furnaces, steel is passed through the
furnace along which the temperature is gradually increased
up to the soak zone. In there, the temperature is kept constant
for thermal and chemical homogeneity of the steel. In the
reheat furnace, being direct fired using natural gas and/or fuel
oil and air, the composition of the "in-furnace" atmosphere,
mainly N2, CO2, H2O and free O2, can vary dramatically.
These variations are functions of the air/fuel ratio, which in
turn, depends on furnace and mill operating conditions.
Hence, reheating under industrial conditions is difficult to
reproduce in the laboratory. As an alternative, oxidation in
such conditions can be approximated from well controlled
laboratory experiments [1].
The subject of high temperature oxidation of steels under
different conditions has been studied extensively. The
dependence of oxidation rates on temperature is well
established and is known to obey an Arrhenius relationship
[2]. In general, oxidation can be classified into three stages.
An initial stage characterized by a linear type of oxidation, a
final stage where oxidation is parabolic and an intermediate
stage where a transition from linear to parabolic mechanism
takes place. For oxidation of different steels in CO2 and H2O
atmospheres, the limiting step was found to be the rate of
dissociation of CO2 or H2O to oxygen and CO or H2 on the
oxide surface [3]-[5]. However, for oxidation in O2
atmospheres, the limiting step during the initial period of
oxidation was found to be the rate of oxygen transport from
the gas phase to the reaction surface [6].
After the oxide
layer reaches a certain thickness, oxidation follows a
parabolic rate law where the rate of oxidation is controlled
by the diffusion of the ionic species and vacancies through the
oxide layer [3], [7], [8].
Oxidation in complex atmospheres such those exist in
industrial reheat furnaces, however, received little attention.
Studies showed that oxidation in multi-component gases can
involve the three mechanisms of oxidation all together [9],
[10].
II. REACTIONS KINETICS
Iron oxidizes to form three well known oxides, namely,
wustite (FeO), magnetite (Fe3O4) and hematite (Fe3O4) in
proportions determined by reaction kinetics, where the
predominant oxide being magnetite In general, the oxidation
process can be expressed as:
n
Manuscript received September 05, 2011; revised January 1, 2012.
H. T. Abuluwefa is with the University of Misurata, Libya.
(phone: 218 92 687-3162; e-mail: [email protected]).
ISBN: 978-988-19251-9-0
ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
W 
  = k xt
 A
(1)
IMECS 2012
Proceedings of the International MultiConference of Engineers and Computer Scientists 2012 Vol II,
IMECS 2012, March 14 - 16, 2012, Hong Kong
Where W and A are the sample weight (g) and surface area
(cm2) respectively, kx is the oxidation rate constant and t is
the oxidation time (seconds). The exponential term n
represents the type of oxidation mechanism. If (n = 1),
oxidation is linear with time, and if (n = 2), the oxidation
process is parabolic.
Intermediate or mixed control
oxidation where n is between 1 and 2 can also take place.
Linear oxidation normally takes place during the initial
oxidation period. During this stage, oxidation is controlled
by the diffusion rate of oxygen atoms from the bulk of the
gas to the reaction surface. Hence, this is the case where
oxidation is a gas phase controlled process. The linear
oxidation can be expressed as:
W 
  = k l .t
 A
(2)
where kl is the linear oxidation rate constant (g/cm2 s). The
linear rate constant in the case of the oxidation of iron in
oxygen-nitrogen gas mixtures is controlled by the rate of
transport of oxygen molecules through the gas phase to the
reaction surface [11]. The linear rate constant for oxidation in
oxygen environment can be expressed as:
Where kp is the parabolic rate constant (g2/cm4. s), and t is the
time in seconds. Assuming that the oxidation product is FexO,
the parabolic rate constant can be expressed as follows [11]:
kp=6
ρ 2FeO M O 2
M
2
G
*
(3)
in which Mo is the molar mass of atomic oxygen (g/g-mol),
kMTC is the mass transfer coefficient (cm/s) and Co2 is the
molar concentration of oxygen in the gas mixture (moles/cm3),
where the superscript G and * refer to the bulk of the gas and
the sample surface, respectively. The mass transfer coefficient
can be obtained by solving the convective mass diffusion
equation across a laminar boundary layer to the surface of a
flat plate, assuming a known bulk gas velocity [12]. The final
equation expressing the mass transfer coefficient, kMTC , is as
follows:
k MTC =
4 DO2
.(Re )1/2 (Sc )1/3
.
3 l
(4)
where Re is the dimensionless Reynolds number (ul/v), Sc is
the dimensionless Schmidt number (v/Do2), MO is the molar
mass of atomic oxygen (g/g-mol), Do2 is the diffusion
coefficient of oxygen in the binary gas mixture (cm2/s), u is the
gas velocity past the sample surface (cm/s), l is the length of
the sample (cm), v is the kinematic gas viscosity (cm2/ s). The
binary gas diffusion coefficient, Do2, and the kinematic gas
viscosity, v, were evaluated using conventional methods [13].
After the oxide layer reaches a certain thickness, between
4 x 10-3 mm and 0.1mm, the outward diffusion of metal ions
and/or the inward diffusion of oxygen ions through the oxide
layer become rate controlling [14]. This diffusion controlled
process is known as parabolic oxidation where the parabolic
rate constant is expressed as:
2
W 
  = kpt
 A
ISBN: 978-988-19251-9-0
ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
(5)
(6)
FeO
where ρFexO is the density of FexO (g/cm3), MO is the
molecular weight of atomic oxygen, MFexO is the molecular
weight of FexO, D*Fe2+ is the iron diffusion coefficient in
iron oxide (cm2/s), yFexO/Fe3O4 and yFe/FexO are the iron ion
vacancy concentrations at the FexO/Fe3O4 and Fe/FexO
boundaries, respectively. The vacancy concentrations are
determined from the wustite/magnetite and iron/wustite phase
equilibria, respectively. Values for the vacancy concentrations
in the temperature range of 800 to 1100oC are presented in the
study by Deich and Oeters [11]. The diffusion coefficient,
D*Fe2+, was determined from studies of iron diffusion using the
decrease in surface activity technique for three artificially
prepared iron oxides [15], and, for wustite was found to obey
the following equation:
*
k l,O2 = M O k MTC (C O2 - C O2 )
. D*Fe2+ ( y FeO/ Fe3O4 - y Fe/FeO )
D Fe2+ = 0.118 e
-124,300
RT
(7)
where R is the gas constant (J/mol. oK) and T is the absolute
temperature. Calculations using the above parabolic relations
were used in studies by Yurek, et al. [16] and Garnaud et al.
[17] to calculate the thickness of two layered scales, FexO and
Fe3O4. Good agreement was found between the calculated and
measured values.
Since oxidation of steel in the reheat furnaces are
extensive, i.e., initial oxidation rates where thin oxide layers
are formed, can be ignored, and hence, parabolic rates can
be considered. In evaluating the parabolic rate constant, the
data collected during the first 10 minutes of oxidation was
omitted. This ensures that data where oxidation is linear
with time are not included in the evaluation of the parabolic
rates.
III. EXPERIMENTS
The experiments were carried out by reacting rectangular
samples with compositions shown in Table I and dimensions
(20mm x 10mm x 10 mm) prepared from a steel billet cross
sections. A tubular vertical furnace consisting of a pure
alumina process tube with an inside diameter of 25 mm was
used. The furnace temperatures were chosen to be 1000oC,
1100oC and 1200oC and controlled within a ±0.2 oC. Each
experiment was performed for 90 minutes. Gases used are
pre-purified nitrogen, extra dry air and carbon dioxide. The
desired gas compositions were attained by adjusting the flow
rate of each gas using high precision flow meters. Weight
changes of the sample during oxidation and furnace
temperature were digitally recorded using a high sensitivity
microbalance. The experimental set up is shown in Fig. 1.
IMECS 2012
Proceedings of the International MultiConference of Engineers and Computer Scientists 2012 Vol II,
IMECS 2012, March 14 - 16, 2012, Hong Kong
TABLE I
CHEMICAL COMPOSITION OF INVERSTIGATED
CARBON STEELS (wt%) x 10-3.
Steel
C
Mn
P
S
Si
Cu
Ni
Cr
Mo
A
370
830
12
18
250
190
440
460
161
B
180
720
4
6
220
70
20
20
5
In industrial reheat furnaces, the composition of the furnace
atmosphere depends on the magnitude of the input air/fuel
ratio used. Calculated gas compositions corresponding to the
combustion of natural gas with maximum, intermediate and
minimum air/fuel ratios are given in Table II.
The
experiments were carried out with furnace atmospheres
corresponding to those given in the table as well as in related
binary gases of O2-N2, CO2-N2 and H2O-N2 leading to the final
gas compositions.
factors defining the shape of the sample weight change curves
during oxidation are the weight gain due to oxygen pick-up
(oxidation) and weight loss due to carbon diffusion out of the
sample surface (decarburization).
If decarburization is
predominant a weight loss is observed, however, if oxidation
is predominant a weight gain is observed
A. Effect of Temperature
Experiments were carried out first in nitrogen-based oxygen
atmospheres in the three temperatures of 1000oC 1100oC and
1200oC where oxygen was maintained at 6 vol.% during these
experiments. Typical samples weight change curves during
oxidation of the two carbon steels are shown in Figs. 2 and 3.
As it is well known, oxidation was highly a function of
temperature, however, some differences of oxidation
behaviour at the three temperatures can be seen. At the lower
temperatures rates of oxidation followed approximately linear
type of oxidation mechanism throughout the oxidation
process, especially at the temperature of 1000oC. This
behaviour is a result of the controlling mechanism of
oxidation. That is, since the oxide layer is very thin and
provided no significant resistance to the transport of
oxygen/iron ion to the reaction surface, the rate of the
chemical reaction at the steel surface is controlling and hence
rates of oxidation exhibited a linear behaviour. For oxidation
at 1200 oC, and after a short period of time lasting about 10
minutes, where oxidation rates followed a linear rate law as
described above, the oxide layer reached a thickness after
which the oxidation process became controlled by the solid
state diffusion of the oxidizing species through the scale layer.
In this case, oxidation was independent of furnace atmosphere
composition. This behavior was observed for the rest of the
experiments where oxidation was conducted at different
atmospheres on the two steels.
Fig. 1. Schematics of the experimental set up used for the oxidation
experiments.
TABLE II
CALCULATED GAS COMPOSITIONS RESULTING FROM THE
COMBUSTION OF NATURAL GAS WITH DIFFERENT AIR/FUEL
RATIOS (vol.%).
Atmosphere
I
II
III
Low
Intermediate
High
Air/gas
Ratio
10
14
37
H2O
O2
CO2
N2
20
15
7
3
5
15
10
8
3
bal.
bal.
bal.
IV. RERSULTS AND DISCUSSION OF RESULTS
Because of the large number of experiments conducted in
this work only few results representative experiments will be
presented. The experiments were performed in single gas
atmosphere, binary gases and finally ternary gases
representing those found in industrial steel reheat furnace
atmospheres shown in Table 2. The purpose for this strategy is
to investigate the effect of single gases and relate that to the
more complex multi-component gases. In general, the two
ISBN: 978-988-19251-9-0
ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Fig. 2. Weight change curves for oxidation of 0.18 wt.%C steel.
IMECS 2012
Proceedings of the International MultiConference of Engineers and Computer Scientists 2012 Vol II,
IMECS 2012, March 14 - 16, 2012, Hong Kong
Fig. 3. Weight change curves for oxidation of 0.32 wt.%C steel.
Fig. 4. Arrhenius plot for oxidation of 0.18 wt.%C steel.
B. Oxidation Rate Constant
The rates of oxidation of the two steels can be best
compared by looking at their oxidation rate constants. The
parabolic rate constants for oxidation of the two steels can be
obtained from the analysis of the weight gain data during
oxidation and are given in Table III. From this data is clear
that rate of oxidation for the two carbon steels at different
temperatures are quite similar and differ only by about 15% at
the most. The similarity in oxidation behaviour can be also
illustrated by comparing the reaction activation energies
obtained from the Arrhenius plots shown in Figs. 4 & 5.
TABLE III
MEASURED PARABOLIC RATE CONSTANTS FOR OXIDATION OF
THE TWO STEELS IN AN ATMOSPHERE OF 6 VOL.% O2
kp (g2/cm4 s)
Temperature, oC
1000
0.18 wt.%C
-8
4.2 x 10
0.32 wt.%C
Fig. 5. Arrhenius plot for oxidation of 0.32 wt.%C steel.
-8
5.1 x 10
1100
-7
2.38x 10
3.75 x 10-7
1200
8.87 x 10-7
1.12 x 10-6
.
For oxidation of the 0.18wt.%C the activation energy was
found to be (238 kJ) and that for oxidation of the 0.32wt.%C
was (242 kJ), both indicative of a parabolic rate type of
oxidation mechanism. From this data the parabolic rate
constant over the temperature range 1000 to 1200oC for the
0.18wt.%C and 0.32wt.%C, respectively, can be expressed as:
k p = 256.21 e
−
k p = 486.8 e
−
238179
RT
242128
RT
 g2 
 4 
 cm s 
(8)
 g2 
 4 
 cm s 
(9)
ISBN: 978-988-19251-9-0
ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
C. Comparison of Oxidation in the Three Atmospheres
The oxidation experiments have shown that in atmospheres
containing 3 and 5% O2 are fairly similar. However, oxidation
in atmosphere III, where the free oxygen content is 15 vol. %,
is different. This difference is seen during the initial period of
oxidation where the rates were much higher. The initial rate
of oxidation is a gas phase transport control and is highly
dependent on oxygen concentration in the gas. In general,
furnace atmospheres containing higher levels of free oxygen,
the
overall oxidation is unaffected by changing the
concentrations of CO2 and H2O.
D. Metallographic Examination
The phase composition of oxide layers after oxidation of the
samples at different conditions was examined using electron
microprobe. It was found that oxidation in atmospheres
containing higher levels of free oxygen produced oxide layers
which consisted of the three phases, wustite FeO, magnetite
Fe3O4 and a very small layer of hematite Fe2O3. However,
IMECS 2012
Proceedings of the International MultiConference of Engineers and Computer Scientists 2012 Vol II,
IMECS 2012, March 14 - 16, 2012, Hong Kong
oxidation in the atmosphere containing 3 vol.% O2 only
wustite and magnetite were present. Shown in Fig. 6 is a
microstructure of 0.72 wt.%C steel which exhibited similar
structure of the two carbon steels being presented in this
article. For comparison purposes, the microstructure of the
oxides produced in an actual reheat furnace was also
examined, Fig. 7. In this case, the microstructure consisted
mainly of wustite with the presence of a thin layer of
magnetite.
rates was the free oxygen. The higher the carbon content the
lower the overall oxidation of the steel, especially at high
temperatures. To control scaling rates of the carbon steels
studied this required control of the level of free oxygen in the
furnace atmosphere.
REFERENCES
[1] H. T. Abuluwefa, R.I.L. Guthrie, and F. Ajersch, (1997), Metallurgical
Transactions A, 28, pp. 643.
[2] L. Himmel, R. F. Mehl, and C. E. Birchenall, (1953), J. Met. Trans.,
AIME, 827.
[3] F. S. Petit and J. B. Wagner, Jr., (1964), Acta Metallurgica, 12, pp. 35.
[4] F. Pettit, R. Yinger and J. B. Wagner, Jr, (1960), Acta Metallurgica, 8,
pp. 617.
[5] W. W. Smeltzer, (1960), Acta Metallurgica, 8, pp. 377.
[6] H. Abuluwefa, R. I. L. Guthrie, and F. Ajersch, (1996), Oxidation of
Metals, 46, 5/6, pp. 423.
[7] C. Wagner, (1933), Z. physik. Chem., B21, pp. 25.
[8] M. H. Davies, M. T. Simand, and C. E. Birchenall, (1951), Journal of
Metals, Transactions of AIME, pp. 889.
[9] H. T. Abuluwefa, R. I. L. Guthrie and F. Ajersch, (1997), Metallurgical
Transactions. A, 28A, pp. 1633.
[10] Hans-Joachim Selenz and Franz Oeters, (1984), Report from the
Institute of Metallurgy (Ferrous Metallurgy) of Berlin Technical
University,
[11] Von J. Deich und F. Oeters, (1973), Werkstoffe und Korrosion, 24,
Heft 5, pp. 365.
Fig. 6. Oxides microstructure of a (similar) 0.72wt.C after oxidation in the
Laboratory.
[12] V. G. Levich, (1962), Physicochemical Hydrodynamics, Prentice
Hall, Englewood Cliffs, N. J. pp. 87.
[13] S. Chapman and T. G. Cowling, (1970), .Mathematical
Theory of Non-Uniform Gases, 3rd Ed., Cambridge
University Press.
[14] F. S. Pettit, R. Yinger, and J. B. Wagner, Jr., (1964), Acta
Metallurgica, 8, pp. 35.
[15] L. Himmel, R. F. Mehl and C. E. Birchanall, (1951), J. Metals,
Transactions AIME, Vol. 3, pp. 889.
[16] G. J. Yurek, J. P. Hirth, and R. A. Rapp, (1974), Oxidation of
Metals, Vol. 8, No.5, pp. 265.
[17] G. Garnaud and R. Rapp, (1977), Oxidation of Metals, Vol. 11,
No.1, pp. 193.
Fig. 7. Oxides microstructure of a 0.72wt.C after oxidation
in the industrial reheat furnace.
V. CONCLUSIONS
This work showed that the overall isothermal oxidation of
high carbon steel in multi-component nitrogen based
atmospheres (O2-CO2-H2O) was oxygen and temperature
dependent. At higher temperature, i.e., 1200oC, oxidation
rates increase with increasing oxygen content up to about 6
vol.%, beyond which no increase in oxidation rates was
observed. Oxidation in atmospheres containing free oxygen
exhibited a parabolic type of oxidation behaviour. However,
oxidation in CO2 and H2O nitrogen based atmospheres
followed a linear rate law and was mainly temperature
dependent. In the quaternary atmospheres, the most
important component contributing to higher initial oxidation
ISBN: 978-988-19251-9-0
ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
IMECS 2012